Six Brilliant Moves and One Great Move in a "Study" with 15 Moves

https://www.chess.com/a/2iKnWoyFCuRkS
Based on the nice blog https://www.chess.com/blog/Mikhail_Golubev/the-old-tricks-7-a-loss-by-a-boss-an-evergreen-brilliance
For some reason the first sack is great instead of brilliant.
But this is my first time seeing a legit game with this many brilliancies, and also White has 0% accuracy while Black 100%. Well, I admit, even though it is possible to get such a position, with lots of pawns promoting for Black, White's pawns mostly getting gobbled up, but still managing to have two left and then promote 2 for a total of 3 queens, it is not likely to happen in a real game...
(Luckily this is not one of those fake brilliant games where every move is brilliant because the engine has depth 0 or something, they are meaningless.)

I am not claiming credit in discovering this, it is very probable someone discovered this before me.

The engine starts out favoring White, which I think is a natural deficiency in all engines where the mate is inevitable but the opposite side has a huge material advantage, especially when the mate requires lots of sacrifice, since they trim their search trees to favor positions that relatively immediately gain material or something...but the engine still finds the win in this case after a while. Could this extend over to engines having problems finding "mate in infinity" on an infinite chessboard while humans can identify some using pattern recognition? (Like this mate but on an infinitely wide chessboard with an infinite number of rooks; also, in that case, we may need to revise the definition of checkmate, but that's another topic.)

Also, does anyone else have a real possible game with more brilliant moves?